SMAUG
Simulating Machine Learning Applications on gem5-Aladdin
|
Network encapsulates all of the information SMAUG will use during execution: the overall computation graph of the model, all the operators and tensors, various housekeeping structures, and simulation information. More...
#include <network.h>
Network encapsulates all of the information SMAUG will use during execution: the overall computation graph of the model, all the operators and tensors, various housekeeping structures, and simulation information.
Classes | |
struct | OperatorInsertion |
Public Member Functions | |
Network (std::string _name) | |
void | addOperator (Operator *op) |
void | addEdge (Operator *src, Operator *dest, TensorIndices indices) |
const OperatorMap & | getOperators () const |
Operator * | getOperator (const std::string &name) |
const Graph & | getGraph () const |
void | dumpDataflowGraph () const |
void | printSummary () const |
bool | validate () const |
OperatorMap::iterator | begin () |
OperatorMap::iterator | end () |
void | setSamplingInfo (const SamplingInfo &_sampling) |
SamplingInfo & | getSamplingInfo () |
Protected Types | |
typedef std::map< std::string, Operator * > | OperatorMap |
Protected Attributes | |
Graph | graph |
The dataflow graph. | |
OperatorMap | operators |
Global map of operator names to their operator objects. | |
SamplingInfo | sampling |
The sampling information of the model. | |
std::string | name |
Name of the model. | |